An experimental investigation on workload limiting methods within ORR policies in a job shop environment

The performances of three workload limiting policies are analysed, and the following are the three objectives: (i) assessing whether the method of workload limiting affects the performances of Order Review and Release strategies; (ii) investigating the performances of the workload limiting methods when the mix imbalance changes; (iii) evaluating the robustness of the workload limiting methods considered. The methodological pattern followed required a simulation model of a dynamic job shop system and all the workload limiting methods are tested by resorting to a fractional factorial experimental design with repetitions. Results coming from simulation campaign show that the 'upper bound only' method is the best performer. Furthermore three different mix imbalance levels have been tested, showing that 'workload balancing' method allows closer performances to those of the 'upper bound only'. Finally 'work-load balancing' is proven to be the most robust, but not all the results are statistically significant for all performance indexes.